The major emphasis of this proposal, which continues the work of the previous proposal, is to devise ways to determine which mutations are involved in adaptive evolution of bacteria as human pathogens. The model system used here is the adaptation of adhesins and other genes of Escherichia coli to extra intestinal infections, particularly urinary tract infections (UTI). The within-clonal genetic diversity of the major uropathogenic serotypes will be used to determine gene loci targeted by pathoadaptive mutations, i.e. gene changes that are selected in the environment when the organism is a pathogen. We will study in detail strains belonging to O18:K1 :H7 serotype - a major uropathogenic clone. Nucleotide polymorphisms within the gene clusters encoding various adhesive fimbria will be determined and used to characterize ancestral/descendent relationships within the clone. Then, two ancestral and two descendant strains will be surveyed for additional mutational changes in up to one megabase of genome by using the newly developed technique, GIRAFF. In GIRAFF, sized fragments from two strains are melted, rehybridized and treated with the mismatch-specific endonuclease, CEL 1, that is capable of cutting mismatched DNA regions with high specificity and sensitivity. The CEL I-specific bands are then identified by Southern blot hybridization using multi-kb DNA probes. Within the 20% of the E. coli genome to be surveyed, about two-dozen synonymous mutations are expected and will be used to date the clone. All other mutations will be analyzed as potential pathoadaptative changes. The genes in which the nonsynonymous mutations are found and the intervening regions where mutations are found will be tested to see if similar mutations are found in the same regions of DNA in six other uropathogenic clones. Those regions commonly found with mutations within each of these other clones will be assumed to important in pathogenesis. The functional effects of these mutations will be investigated. A subset of these pathoadaptive loci will be sequenced in our collection of 125 E. coli strains, which includes both commensal and pathogenic strains, to see if our new analytic technique, zonal analysis, will confirm that these changes are pathoadaptive. If so, this approach can be used to discover pathoadaptive loci from the many expected to be sequenced genomes of E. coli.